THE ROLE OF ARTIFICIAL INTELLIGENCE IN INVESTMENT

MANAGEMENT: ENHANCING DECISION-MAKING, EFFICIENCY,

AND ALPHA GENERATION
Dr. Gaurav Jangra1 Monika Jangra2
1
Associate Professor, Management, Chandigarh University, [email protected]
2
Assistant Professor, Computer Applications, Chandigarh University, [email protected]

1. INTRODUCTION:
The financial industry has witnessed a revolutionary change in advent of AI
technologies for portfolio management. A.I.'s data-processing, pattern-recognizing, and
prediction-making capabilities has significantly impacted decision-making processes and
investment strategies.
Incorporating AI and automation into financial investment adds a new dimension, as
these two comments demonstrate. In traditional quantitative financial investing, humans were
responsible for coming up with investment strategies (Ou & Penman, 1989); (Holthausen &
Larcker, 1992). However, with AI, the machine learns from its own performance and changes its
strategy based on market conditions. This frees up humans to concentrate on building the right
machine, or improving its learning process. As a result of this shift in focus, developers' work
is now mainly assessed for the AI's resilience, persistence, and consistency in its training
process rather than its performance (yield). Instead of creating a strategy, developers create a
machine that can generate its own strategies. Creating strategies that can change with the times
is the end goal of AI-based tech investment. However, if the investment plan ends up
underperforming as a result of the machine learning process, then who is responsible? Issues
with comprehending, evaluating, and controlling the ethical and technical soundness of the
programmer's work and its contribution to wealth creation may arise when an AI developer is
faced with a trade-off between short-term and long-term portfolio performance (Beccalli
Elenaand Elliot, 2020)

(Keding & Meissner, 2021) study show that AI-based recommendations influence how people
make decisions and distort their views of the consequences of such decisions. The authors
conducted a choice experiment using vignettes to gauge how 150 top managers felt about AI-
enhanced decision-making on a personal level. Usage of AI enabled advising process has a
good effect on decision-making behaviour and quality of decisions, according to their study.
Their research goes on to demonstrate that over-reliance on AI-enhanced decision-making is a
greater degree of trust in the advisor, along with the expectation of a more structured procedure.
There is a one-of-a-kind opportunity for expansion in the Islamic financial services market due
to the rising number of Muslims around the world (Irfan et al., 2023).

2. REVIEW OF LITERATURE

(Beccalli et al., 2020) seeks to investigate ethical concerns related to the utilization of
Artificial Intelligence (AI) in the financial management of portfolios, and their managerial
consequences. Traditional quantitative investing in the past involved making portfolio
allocation decisions based on a well-outlined investment strategy. Financial portfolio managers
create and implement investment strategies to optimize expected returns for clients' portfolios.
AI-enhanced algorithms now allow intelligent computers to autonomously adjust and improve
investing strategies by analyzing historical data. Artificial intelligence can have a substantial
impact on the outcomes of portfolio management methods, leading to ethical problems around
human versus machine duty, accountability, and risk. Managers need to implement new
methods for monitoring performance, evaluating competence, and allocating incentives while
overseeing AI software developers in this field.

The ultimate goal of artificial intelligence (AI) is to augment or even completely replace
human intelligence in almost every area where it is now used. Many different fields are making
use of AI as a result of both internal and external factors, such as technical progress. Among
these, the use of AI in the financial sector has bright prospects. This article outlines the
fundamental methods and specific situations used in artificial intelligence (AI) applications
across several domains, with a focus on the financial sector. Using spectral clustering (SC)
linked to a stock complex network as an example, this article also builds a portfolio using AI
in the field of portfolio management. In this scenario, the AI-constructed portfolio beats the
more conventional ones(Zhang & Chen, 2017).

(Abdelazim & Wahba, 2006) The issue of portfolio selection and management is one that
this study seeks to go into. Working from a collection of assets (equities), modern portfolio
theory and the Markowitz efficient frontier determine the ideal weight combination for a risky
portfolio that is on the efficient frontier. Step one of the present study is to determine which
assets should be chosen initially from a set of available assets; step two is to attempt to forecast
returns in order to make better use of the Markowitz efficient frontier. Utilising artificial
intelligence (AI) techniques to discover a quantitative and systematic approach to building an
ideal portfolio was driven by the widespread use of these techniques in numerous areas of
finance. A number of artificial intelligence methods are finding effective solutions to difficult
optimisation issues, and genetic algorithms (GAs) are among them. In this study, GAs are used
to choose the best portfolio by maximising a composite objective function that takes into
account the return, risk, and cross-correlation of each investor's assets. GAs are evaluated on
two stock markets: one in the United States, with 40 businesses representing that market, and
another in Egypt, with 37 companies representing that market that are currently trading. During
both the training and testing periods, the genetic algorithm-generated ideal portfolio
outperformed the market index in terms of risk-adjusted returns (TARGET). Compared to the
more traditional method of averaging past returns, the neural networks approach yields more
accurate predictions of future yields. The ideally chosen portfolio, which was overseen weekly
by neural networks, was discovered to be able to produce positive returns utilising the
Markowitz efficient frontier, even during bad market seasons. A combination of genetic
algorithms and neural networks, the suggested AI strategy, has shown promise in the study's
active portfolio management and selection applications.

An intelligent system that aspires to assist in investment decision-making is Intelligent

Investor. In instance, the system facilitates the management processes, portfolio construction,
and stock selection procedure for the Athens Stock Exchange. The goal of implementing these
systems is to optimise their response to the complex and ever-changing financial markets by
utilising AI technologies in addition to the benefits of Decision Support Systems (models,
interactive decision-making procedure, preferential reasoning). The suggested setup makes use
of AI tools (expert systems) and multi-criteria analysis techniques. All the most up-to-date
methods for managing a portfolio are part of this integrated approach, including market
psychology, technical analysis, and fundamental analysis. In order to build an ASE stock
portfolio that is uniquely suited to the user-investor's profile, the system gives a comprehensive
and sufficiently substantiated investment proposal (Samaras & Matsatsinis, 2004).

More and more, financial markets are utilising AI systems, thanks to the remarkable
advancements in this field over the last decade. The use of algorithms and automated trading
systems has been more common among investing firms and hedge funds in recent years.
Institutional and individual investors alike owe their success in the capital markets to their
lightning-fast decision-making and execution processes. By discovering investment
opportunities and timing entrance and exit of trading, algorithmic trading utilising machine
learning algorithms has improved investor performance. In addition to identifying the most
promising market prospects, this research aims to improve portfolio performance by
developing a smart and fully automated trading system that investors may use to their
advantage. The current investigation is structured into four distinct phases. First, adjusting the
settings of technical indicators; second, determining if the market is trending or not; third, using
the signals to issue a clear signal (buy, sell, or hold); and lastly, rebalancing the portfolio. A
genetic algorithm, fuzzy logic, an artificial neural network, and a traditional portfolio
optimisation model each carry out one of these four processes. The results demonstrate that the
suggested model outperforms the TEDPIX and other mutual funds over the same time period
in terms of generating higher returns with lower risk (Haddadian et al., 2022).

Soft computing techniques, such as genetic algorithms and artificial neural networks,
have been used by numerous researchers in recent decades to solve mathematical issues
(Goldberg, 1998). A wide range of issues were effectively addressed using these techniques.
Most smart systems made use of soft-computing techniques. Meanwhile, financial difficulties
are one of its applications, and many scholars are looking into how soft-computing technics
(the preferred term for "techniques" in formal contexts) might be applied in this area (Baba et
al., 2001).

Many trading methods are based on conventional technical analysis, which is widely
used in markets around the world. It often takes the shape of various trading rules and focuses
on where trends' prices have reached their highest and lowest points. Stock trading systems that
rely solely on technical guidelines to make decisions will, alas, no longer be applicable because
they rely on traders' personal experiences. This is the most significant disadvantage of
employing this analysis method. That some criteria form the basis of the technical rules is the
fundamental issue. The success of an investor depends on the accuracy and wisdom of the
technical factors that are used to make trading decisions. Another issue is that decision-making
becomes more complicated when various technical rules are used at the same time to provide
contradicting trade signals.

In a 1994 study, Refenes et al. examined the efficacy of neural network and regression
models in predicting stock market behaviour. To forecast stock prices, the researchers in this
study used neural networks instead of traditional statistical methods. This study's findings
demonstrated that, compared to statistical methods, neural networks produced superior models.
A system that could anticipate the large, short-term shifts in stock price was developed by Tan
et al. (1995). The data was pre-processed before the neural network, which accurately predicted
trading positions, was modelled.
 In order to forecast which stocks to purchase, sell, and hold, Kuo et al. (2001) created
a method for the stock market. The created system's defining feature was its capacity to quantify
qualitative characteristics that impeded stock price prediction. In 1998, these same-titled
researchers released a work that did not use genetic algorithms.
In order to utilise expert opinions in predicting stock values, the researchers in the cited work
used a questionnaire utilising the Fuzzy Delphi Method (FDM).

Research by Yim (2002) contrasted ARMA and GARCH models, two popular
econometric tools. The outcomes demonstrated that neural networks outperformed the
traditional ARMA and GARCH models. When it came to modelling and short-term prediction
of the Iranian exchange rate, Nouri et al. (2014) compared ARIMA with artificial neural
networks (ANN). Their findings showed that the neural network they used outperformed the
ARIMA model in terms of prediction power.

Foroughi et al. (2014) conducted research that used a mix of artificial neural networks
and an optimisation technique based on particle aggregative movement to forecast the
profitability of each stock. Their findings demonstrated that the multivariate model accurately
forecasted the profit of every stock with a precision of 92% and the univariate model with a
precision of 78%.

Using fuzzy logic, Souto-Mairo (2011) attempted to forecast the future movement of
the Brazil Stock Price Index; in the end, their forecasts were deemed appropriate. In 2011,
Alejandro Rodríguez and colleagues utilised neural networks to enhance indications used in
technical analysis. In order to better anticipate when a market will turn over, this study
employed a genetic algorithm to refine the technical analysis prediction parameters and
strengthen the ENS networks with refined parameters. Employed strategies outperform buy-
and-hold strategies, according to research (Lin et al., 2011).

To maximise a portfolio of Tehran Stock Exchange mutual funds, Rahnamay et al.

(2015) used a genetic algorithm strategy. The outcomes suggested that the genetic algorithm
would be a better choice for picking the genetic algorithm-generated portfolios. When
researching the best time to enter technical stock trading, Tehrani and Abbasion (2008) relied
solely on a neural network technique. Their findings showed that the strategy worked best
during bear markets. The suggested trading strategy was statistically indistinguishable from the
hold and purchase strategy during bull markets.
 The top 1% of stock price changes across 10-second trading periods are considered
extraordinary price movements, and Brogaard et al. (2018) study high-frequency trading in
relation to these extreme price fluctuations. According to the authors, the overall impact of
liquidity demand and supply in high-frequency trading is inversely related to the direction of
price fluctuations.

Trading rules were generated by Michell et al. (2020) using a Fuzzy inference system
and an embedded STPG technique. Their methodology, FISTGP, utilised a combination of
fuzzy inference systems (FIS) and strongly typed genetic programming (STGP) to produce
trading rules for the US stock market. Taking excess returns into account, the suggested model
beats the Buy and Hold (B&H) approach during the testing period by 28.62% while carrying
nearly the same amount of risk (1.28% greater).

Using a STGP approach, Manahov (2018) investigated the impact of high-frequency

data on the market and found that it hurts institutional traders millisecond by millisecond. In
order to spot trends in the bitcoin market, Ha and Moon (2018) suggested GP. Applying market
signals to boost investment return is the main emphasis of the study. The results of these tests
demonstrate that GP is capable of developing effective trading rules.

Additionally, using data collected overnight, Rastegar and Sedaghatipour (2019)

devised an algorithmic trading system for futures contracts involving gold coins. The Signal
8 trading strategy leverages technical analysis, which works well in markets with both
buying and selling positions. By maximising the two functions of return and conditional
value at risk (CVaR), a trading system has been constructed using the MOPSO algorithm.
The ideal profit and loss limit for the futures contract has also been defined, completing the
risk management system. The results demonstrate that compared to other competing
strategies, such buying and holding or selling and holding, the developed trading strategy
offers a higher return-to-risk ratio.

An attempt at developing a technical analysis-based intelligent trading system has been

made by Baky et al. (2021). Make use of machine learning and meta-heuristic methods.
They optimised the parameters of technical indicators using a genetic algorithm, and then
used an artificial neural network to transform the signals into buy or sell orders. The
research-based trading strategy differs significantly from the buy-and-hold approach to
stock investing, according to the data.

3. PORTFOLIO MANAGEMENT
 One popular field of study is methods, primarily algorithms, for finding the best way to
distribute assets across different types of investments. The financial market is a complicated
nonlinear dynamic system, and ANNs are the most well-known of these algorithms because of
unstructured characteristic of investor decision-making and the inherent uncertainty of the
market environment. One surefire way to handle portfolio management is with ANNs, or
Artificial Neural Networks, which can accurately grasp the nonlinear characteristic. ANNs
have demonstrated superior performance compared to conventional models in a variety of
areas, including mortgage, stock, real estate, and accountancy(Holthausen & Larcker, 1992).
Hybrid AI approach to Portfolio Management
Fig.1. Construction and Training of the Model

Source: Haddadian et al., 2022

Fig.2. Implementation and evaluation

 Source: Haddadian et al., 2022

3.1 Quantitative Analysis:

Finding quantifiable trends in market data over time using machine learning algorithms
to guide investing choices. Enhancing portfolio optimization through AI-driven algorithms that
adapt to changing market conditions. By concentrating on prediction via ensemble learning and
being able to grasp nonlinear patterns, active portfolio management can reap the benefits of
machine learning (ML) techniques. With the advent of Industry 5.0 comes the need for
sophisticated tools to help financial institutions make sense of the myriad of investment,
macroeconomic, and credit-related decisions that will arise(Irfan et al., 2024).
Various stages of the investment process can benefit from ML methods, including as
signal generation, portfolio construction, and trade execution. The industry is anticipating
reinforcement learning to have a larger impact. In terms of actual investment returns, exchange-
traded funds that use machine learning have shown varied results (Söhnke et al., 2021)
3.2 Risk Management:
Implementing AI models for real-time risk assessment, identifying potential market
risks, and dynamically adjusting portfolios to mitigate exposure. Enhancing stress testing and
scenario analysis through advanced AI simulations. According to Woodall (2017), there are
several real-world examples of machine learning being used to validate models. Among these
is Nataxis, a French investment firm that runs over 3 million simulations nightly using
unsupervised learning to find new asset-connection patterns and investigate any that show
'errant' patterns compared to average estimates. Additionally, Woodall brings attention to the
fact that Nomura use machine learning to keep tabs on internal trading, ensuring that
inappropriate assets are not being utilized in trading models. The company yields.io offers a
fascinating modern take on model risk management by using AI and ML to power real-time
model monitoring, deviation testing, and validation. (Irfan, 2021) demonstrates the evolving
connection between domestic Shariah Indices of GCC states, which is crucial for investors to
manage their risk and for policymakers to pre-emptively address the potential financial decline
in their countries.

Fig. 3. Use of AI/ML in Investment Management

Source: Refinitiv AI/ML survey August 2020

4. ALPHA GENERATION:
Saizhuo et al. (2023) suggested alpha mining to solve these problems; it improves the
efficacy and efficiency of alpha research by strengthening human-AI connection. They offer
the design of an interactive alpha mining system, Alpha-GPT, based on this novel paradigm.
The method uses LLM to bridge the gap between quantitative researchers and alpha search.
Three major benefits characterize alpha-GPT. One of its primary functions is to understand and
articulate the trading concepts expressed by its users. Secondly, Alpha-GPT can swiftly
summarize top-performing alphas in natural language for ease of comprehension. Last but not
least, after the user has made their suggestions, the model will automatically incorporate them
into subsequent rounds of alpha mining to improve the alpha search.(Irfan et al., 2021)
4.1 Predictive Analytics:
Harnessing machine learning models to forecast market trends, stock movements, and
economic indicators. Identifying alpha-generating opportunities by leveraging AI for predictive
stock selection.
4.2 Sentiment Analysis:
Using NLP to sift through textual data such as news articles, social media posts, and
more in order to glean insights on market sentiment. Incorporating sentiment analysis into
investment strategies to gain insights into market perception and investor behavior. The book
by Irfan et al., 2023) emphasis on intricate financial applications including financial risk
management in big data settings, this book presents fresh research on the functions of AI and
ML algorithms in the financial sector.
5. OPERATIONAL EFFICIENCY
One game-changing innovation is AI, or artificial intelligence. Businesses in many
kinds of industries are starting to use AI to streamline their processes, and many managers have
already put a lot of money into AI in the hopes that it will enhance their operations (Lui et al.,
2022). In the banking sector, for example, Castellanos (2018) notes that Bank of America has
employed a AI chatbot named Erica to assist clients with simple banking chores, while Bank
of New York Mellon Corp has deployed over 220 "bots" to do repetitious jobs. By cutting 22
percent of their companies' operational costs, Banks might save more than $1 trillion by 2030
with the help of AI., according to analysts (Joyce 2018). Additionally, according to research by
McKinsey Global Institute, artificial intelligence has the potential to add $13 trillion to the
world's annual economic production by 2030 (Bughin et al. 2018).
Statista (2016) predicted that by 2025, AI-powered business apps may generate $31.2
billion in revenue, up from $1.62 billion in 2018. Not only that, but 84% of businesses think
AI will give them a competitive advantage, and 75% think it will open doors to new prospects.
Further, research from Fortune (Murry 2017) shows that over 80% of Fortune 500 CEOs
consider AI to be very important for the future of their companies.
5.1 Automation of Routine Tasks:
 Streamlining routine tasks such as data entry, reconciliation, and reporting through AI-
driven automation, allowing investment professionals to focus on higher-value activities.
Improving operational efficiency by reducing manual errors and minimizing processing times.
5.2 Fraud Detection and Compliance:
Implementing AI-based fraud detection systems to identify irregularities and enhance
the overall security of investment processes. Ensuring compliance with regulatory
requirements through AI-driven monitoring and reporting.
5. CHALLENGES AND CONSIDERATIONS:
 Addressing challenges related to the interpretability of complex AI models to enhance
transparency and build trust among investors and regulators.
 Mitigating risks associated with sensitive financial data by implementing robust
cybersecurity measures.
 Anticipating the evolution of AI technologies in investment management, including the
integration of explainable AI and increased collaboration between human expertise and
machine intelligence.
CONCLUSION
The role of AI in investment management is evolving rapidly, reshaping traditional
practices and unlocking new opportunities for alpha generation and operational efficiency.
While challenges exist, the continued advancement of AI technologies promises to enhance
decision-making processes, mitigate risks, and drive innovation in the dynamic landscape of
investment management. As AI continues to mature, investment professionals must adapt to
leverage its full potential in navigating the complexities of financial markets.

REFERENCES:
Abdelazim, H. Y., & Wahba, K. (2006). An artificial intelligence approach to portfolio
selection and management. International Journal of Financial Services
Management, 1(2–3), 243–254. https://doi.org/10.1504/IJFSM.2006.009629

Alejandro Rodríguez et al. (2011). Using neural networks to improve trading systems
based on technical analysis by means of the RSI financial indicator, Expert Systems
with Applications. 38(9):11489-11500. DOI: 10.1016/j.eswa.2011.03.023

Baba, N., et al. (2001). Knowledge-Based Decision Support Systems for Dealing Nikkei-
225 by Soft Computing Techniques. in Knowledge-Based Intelligent Information
 Engineering Systems and Allied Technologies KES. 2001: IOS Press. The
Netherlands. DOI:https://doi.org/10.1007/11552413

Baky, M. Haddadian, H. Zomorodian, G. (2021). An Algorithmic Trading system Based

on Machine Learning in Tehran Stock Exchange. Advances in Mathematical Finance
& Applications. 6(3), 653-669. DOI: 10.22034/AMFA.2020.1894049.1380

Beccalli, E., Elliot, V., & Virili, F. (2020). Artificial Intelligence and Ethics in Portfolio
Management. In R. Agrifoglio, R. Lamboglia, D. Mancini, & F. Ricciardi (Eds.),
Digital Business Transformation, 19–30. Springer International Publishing.

Beccalli Elena and Elliot, V. and V. F. (2020). Artificial Intelligence and Ethics in Portfolio
Management. In R. and M. D. and R. F. Agrifoglio Rocco and Lamboglia (Ed.),
Digital Business Transformation, 19–30. Springer International Publishing.

Dariush Forougi et al. (2014). Earnings Per Share Forecast: the Combination of Artificial
Neural Networks and Particle Swarm Optimization Algorithm, Investment Science.
6, 63-82. DOI: http://jik.srbiau.ac.ir/article_7480.html

Goldberg, D. (1998). Approximation by superposition of a sigmoidal function. Journal of

Mathematics of Control Signals & Systems, 2, 303-314.
DOI:https://doi.org/10.1007/BF02551274

Haddadian, H., Baky Haskuee, M., & Zomorodian, G. (2022). A Hybrid Artificial
Intelligence Approach to Portfolio Management. Iranian Journal of Finance, 6(1),
1–27. https://doi.org/10.30699/ijf.2021.287131.1237

Ha, S. Moon, B. (2018). Finding attractive technical patterns in Crypto-currency markets.

Memetic Computing. vol. 10, no. 3, pp. 301– 306.

Holthausen, R. W., & Larcker, D. F. (1992). The prediction of stock returns using financial
statement information. Journal of Accounting and Economics, 15(2–3), 373–411.
https://EconPapers.repec.org/RePEc:eee:jaecon:v:15:y:1992:i:2-3:p:373-411

Irfan, M. (2021). Do Shariah Indices converge? Evidence from Gulf Cooperation Council
countries. International Journal of Business Excellence, 23(2), 251–269.
https://doi.org/10.1504/IJBEX.2021.113448
Irfan, M., Elhoseny, M., Kassim, S., & Metawa, N. (Eds.). (2023). Advanced Machine
Learning Algorithms for Complex Financial Applications. IGI Global.
https://doi.org/10.4018/978-1-6684-4483-2

Irfan, M., Elmogy, M., Shabri Abd. Majid, M., & El-Sappagh, S. (Eds.). (2024). The
Impact of AI Innovation on Financial Sectors in the Era of Industry 5.0. IGI Global.
https://doi.org/10.4018/979-8-3693-0082-4

Irfan, M., Kadry, S., Sharif , M., & Ullah Khan, H. (2023). Fintech Applications in Islamic
Finance: AI, Machine Learning, and Blockchain Techniques. IGI Global.
doi:10.4018/979-8-3693-1038-0

Irfan, M., Kassim, S., Shaikh, H., Kumar D And, M., & Jhamnani, R. (2021). Do Volatility
and Regime Switching Affect Sustainable Indices Evidence from Global Stock
Markets based on Inductive Approach of GARCH Family. In International Journal
of Economics and Management Journal homepage (Vol. 15, Issue 2).
http://www.ijem.upm.edu.my

J. Brogaard, et al. (2018). High-frequency trading and extreme price movements. Finance.
Econ. 128 (2), pp. 253-265 DOI:10.1016/j.jfineco.2018.02.002

Juliana Yim. (2002). A Comparison of Neural Networks with Time Series Models for
Forecasting Returns on a Stock Market Index. Applications of Artificial Intelligence
and Expert Systems. IEA/AIE, Cairns, Australia, June 17-20, 2002, Proceedings, 25-
35. DOI: https://doi.org/10.1007/3-540-48035-8_4

Kuo, R. J., Chen, C., & Hwang, Y. (2001). An intelligent stock trading decision support
system through integration of genetic algorithm based fuzzy neural network and
artificial neural network. Fuzzy Sets and Systems. 118, 21–45. DOI:
https://doi.org/10.1016/S0165-0114(98)00399-6

Keding, C., & Meissner, P. (2021). Managerial overreliance on AI-augmented decision-

making processes: How the use of AI-based advisory systems shapes choice behavior
in R&D investment decisions. Technological Forecasting and Social Change, 171,
120970. https://doi.org/https://doi.org/10.1016/j.techfore.2021.120970
Lin, X. Yang, Z. Song, Y. (2011). Intelligent stock trading system based on improved
technical analysis and Echo State Network. Expert Systems with Applications. 34,
620-627. DOI: https://doi.org/10.1016/j.eswa.2011.03.001

Lui, A. K. H., Lee, M. C. M., & Ngai, E. W. T. (2022). Impact of artificial intelligence
investment on firm value. Annals of Operations Research, 308(1), 373–388.
https://doi.org/10.1007/s10479-020-03862-8

Michel, K. Kristjanpoller, W. (2020). Strongly-typed genetic programming and fuzzy

inference system: An embedded approach to model and generate trading rules.
Journal of Applied soft computing. 90, 6-25. DOI:
https://doi.org/10.1016/j.asoc.2020.106169

Manahov, A (2018). The rise of the machines in commodities markets: new evidence
obtained using strongly typed genetic programming. Annals of Operations Research.
260 (2), 321–352, 2018.

N. Baba, T. Kawachi, T. Nomura, Y. Sakatani, (2004), Utilization of NNs & Gas for
improving the traditional technical analysis in the financial market. SICE Annual
Conference, 2(2), 1409-1412.

Nouri, A, et al. (2014). Comparing the performance Of Artificial Neural Networks (ANN)
and Auto-Regressive Moving Average (ARIMA) Model in Modeling and
Forecasting Short-term Exchange Rate Trend in Iran. Investment Science Journal.
10, 85-100. DOI: http://jik.srbiau.ac.ir/article_7607.html

Ou, J. A., & Penman, S. H. (1989). Financial statement analysis and the prediction of stock
returns. Journal of Accounting and Economics, 11(4), 295–329.
https://doi.org/https://doi.org/10.1016/0165-4101(89)90017-7

Rahnamay Roodposhty et al. (2015). Optimization of portfolio Constituted from mutual

funds of Tehran stock exchange using genetic algorithm. Investment Science. 12,
217-232. DOI: http://jik.srbiau.ac.ir/article_7672.html

Rastegar, MA. Sedaghatipour, A. (2018). Algorithmic Trading System for the future
contract of gold coin based on intra-day data. Science of investment. 7 (28) 49-67.

Refenes A. N., Zapranis A., & Francis G. (1994). Stock Ranking: Neural Networks vs
multiple linear regression, IEEE. DOI: https://doi.org/10.1109/ICNN.1993.298765
Samaras, G. D., & Matsatsinis, N. F. (2004). Intelligent Investor: An intelligent decision
support system for portfolio management. Operational Research, 4(3), 357–371.
https://doi.org/10.1007/BF02944152

Souto-Mairo, (2011). S&P 500 Index Direction Forecasting from 1976 to 2010: A Fuzzy
System Approach, The International Journal of Digital Accounting Research. 111-
134. DOI: 10.4192/1577-8517-v11_6

Tan, H. Prokhorov, D. Wunsch, D. (1995). Probabilistic and time-delay neural network

techniques for conservative short-term stock trend prediction. in Proc. World Congr.
Neural Networks, https://doi.org/10.1109/72.728395

Tehrani, R. Abbasion, V. (2008). Application of artificial neural networks in Stock Market

Timing: A Technical Analysis Approach. Journal of Economic Research. 8(1), 151-
178. URL: http://ecor.modares.ac.ir/article-18-10926 fa.html

Zhang, X., & Chen, Y. (2017). An Artificial Intelligence Application in Portfolio

Management. Proceedings of the International Conference on Transformations and
Innovations in Management (ICTIM 2017), 775–793. https://doi.org/10.2991/ictim-
17.2017.60